Recent evidence suggests that a family of nerve growth factors exist, each effective for a certain population of neurons. Human fibroblasts make a NGF similar to mouse submaxillary NGF in both its immunoreactive and biological properties. Mouse brain contains a factor which is NGF-like by immunoassay but has no biological activity. This factor increases in the cerebella of the pcd mutant mouse as the Purkinje cells die out and astrocytes proliferate. The mRNA for this factor appears to hybridize with mouse Beta-NGF cDNA and is increased in pcd cerebellum. Entorhinal cortex lesions in rat stimulate production of a factor which can be assayed by its ability to support survival of chick embryo sensory and sympathetic neurons. The mRNA for the factor can be assayed by in vitro translation in an oocyte system.